* doc/lispref/variables.texi (Adding Generalized Variables):
Document it.
* lisp/emacs-lisp/bytecomp.el (byte-compile-warn-obsolete): Alter
the interface so that it can also be used by generalized variable
warnings.
(byte-compile-function-warn): Adjust caller.
(byte-compile-check-variable): Adjust caller.
* lisp/emacs-lisp/gv.el (gv-get): Warn about obsolete generalized
variables (bug#49730).
(make-obsolete-generalized-variable): New function.
'User variables' were made obsolete in Emacs 24 along with
user-variable-p; the sign of the position in (#$ . POS) hasn't
mattered since.
* lisp/emacs-lisp/bytecomp.el (byte-compile-output-docform):
Don't emit negative position when doc string starts with `*`.
* src/lread.c (get_lazy_string): Explain.
* lisp/emacs-lisp/bytecomp.el (byte-compile-output-docform): Don't
insert a \n before the #@ docstrings since make-docfile doesn't scan
.elc files any more.
* src/editfns.c (narrow_to_region_internal): New function, which
contains the body previously in 'Fnarrow_to_region' but accepts
a third argument.
(Fnarrow_to_region): Use the new function. Update the docstring.
(Fwiden): Update the docstring.
* src/lisp.h: Prototype of the new function.
* src/xdisp.c (handle_fontified_prop): Use the new function instead
of 'Fnarrow_to_region'.
* src/process.c (Finternal_default_process_filter):
* src/lread.c (readevalloop): Remove the third argument to
'Fnarrow_to_region'.
* src/bytecode.c (exec_byte_code):
* lisp/emacs-lisp/comp.el (comp-limplify-lap-inst):
* lisp/emacs-lisp/bytecomp.el: Restore the statu quo ante.
* etc/NEWS: Remove the entry about the new optional argument.
* doc/lispref/positions.texi (Narrowing): Update the documentation.
This is a cosmetic change only; there is no change in behaviour.
* lisp/emacs-lisp/bytecomp.el:
* src/bytecode.c (BYTE_CODES, exec_byte_code):
Update and/or remove incorrect, outdated or useless comments.
Clarify. Reorder where appropriate. Rename Bsave_current_buffer to
Bsave_current_buffer_OBSOLETE and Bsave_current_buffer_1 to
Bsave_current_buffer, reflecting the state since 1996.
* lisp/emacs-lisp/bytecomp.el (byte-compile-file): Document behavior
if 'no-byte-compile' is set.
* test/lisp/emacs-lisp/bytecomp-tests.el
(byte-compile-file/no-byte-compile): New unit test.
* test/lisp/emacs-lisp/bytecomp-resources/no-byte-compile.el: New test
file.
* lisp/emacs-lisp/bytecomp.el
(byte-compile-file-form-custom-declare-variable): We already warn
about missing groups from byte-compile-normal-call, so this would
be a double warning.
(custom-declare-face)
(byte-compile-file-form-custom-declare-face): Add doc string
checking for defface.
* lisp/emacs-lisp/bytecomp.el (byte-compile-warning-prefix): Replace a wrong
'or' form involving OFFSET with simply OFFSET. This prevents OFFSET from
possibly being nil in the first branch of the containing `if' form.
* lisp/Makefile.in (BYTE_COMPILE_FLAGS): Enable all byte
compilation warnings.
* lisp/emacs-lisp/bytecomp.el (byte-compile-warning-types): Add
docstrings-non-ascii-quotes and document new semantics for `all'
and t.
(byte-compile--emacs-build-warning-types): New constant.
(byte-compile-warning-enabled-p): Implement the new semantics.
(byte-compile-docstring-style-warn): Reinstate the Unicode quote
warning.
* lisp/emacs-lisp/bytecomp.el (byte-compile-keep-pending):
Allow bytecode split between all kinds of top-level forms, not just
those with chunk handlers, to prevent individual chunks from growing
too large. In particular this helps compilation of
package-quickstart.el.
* lisp/emacs-lisp/byte-opt.el (byte-optimize-form-code-walker)
(byte-optimize-let-form, byte-optimize-letX):
* lisp/emacs-lisp/bytecomp.el (byte-compile-unwind-protect):
Simplify source optimisation and codegen code that can now rely on
normalised let/let* and unwind-protect forms.
Make cconv work for dynamically bound code and always run it.
This allows later stages to benefit from transformations and
normalisations in cconv.
* lisp/emacs-lisp/bytecomp.el (byte-compile-preprocess): Always run
cconv.
* lisp/emacs-lisp/cconv.el (cconv--analyze-function)
(cconv-analyze-form): In dynbound code, treat all variable bindings
as dynamic (lambda, let, let* and condition-case).
Early normalisation of setq during macroexpand-all allows later
stages, cconv, byte-opt and codegen, to be simplified and duplicated
checks to be eliminated.
* lisp/emacs-lisp/macroexp.el (macroexp--expand-all):
Normalise all setq forms to a sequence of (setq VAR EXPR).
Emit warnings if necessary.
* lisp/emacs-lisp/cconv.el (cconv-convert, cconv-analyze-form):
* lisp/emacs-lisp/byte-opt.el (byte-optimize-form-code-walker):
* lisp/emacs-lisp/bytecomp.el (byte-compile-setq):
Simplify.
* test/lisp/emacs-lisp/bytecomp-tests.el: Adapt and add tests.
* test/lisp/emacs-lisp/bytecomp-resources/warn-variable-setq-nonvariable.el;
* test/lisp/emacs-lisp/bytecomp-resources/warn-variable-setq-odd.el:
New files.
* lisp/emacs-lisp/bytecomp.el (byte-compile--reify-function):
Don't transpose doc string and interactive spec, which must come
in this order.
* test/lisp/emacs-lisp/bytecomp-tests.el
(bytecomp-function-attributes): New test.
* lisp/emacs-lisp/bytecomp.el (byte-compile--first-symbol-with-pos):
Traverse record literals as well as vectors. Either is rather
pointless but there were some strong feelings about it.
The bytecode interpreter can't directly call special forms, so
the byte-compiler usually converts special forms into some sequence of
byte codes (basically, providing a duplicate definition of the special
form). There are still two exceptions to this: `defconst` and `defvar`,
where the compiler instead generates a convoluted chunk of code like:
(funcall '(lambda (x) (defvar <sym> x <doc>)) <value>)
where the quote makes sure we keep the function non-compiled, so as
to end up running the special form at run time.
Get rid of this workaround by introducing `defvar-1` and `defconst-1`
which provide a *functional* interface to the functionality of the
corresponding special form.
* src/eval.c (defvar, Fdefvar_1, Fdefconst_1): New functions, extracted from
`Fdefvar` and `Fdefconst`.
(Fdefvar, Fdefconst): Use them.
(syms_of_eval): `defsubr` the new functions.
* lisp/emacs-lisp/bytecomp.el (byte-compile-tmp-var): Delete const.
(byte-compile-defvar): Simplify using the new functions.
* doc/lispref/variables.texi (Defining Variables): Adjust the doc of
`defvar` to reflect the actual semantics implemented.
* lisp/emacs-lisp/bytecomp.el (byte-compile--first-symbol-with-pos)
(byte-compile--warning-source-offset):
Remove recursion for cdr-traversal of lists, and optimise (bug#55414).
* lisp/emacs-lisp/bytecomp.el
(byte-compile-docstring-length-warn): Made into obsolete alias.
(byte-compile-docstring-style-warn): Also warn about other
stylistic issues.
(byte-compile-file-form-autoload, byte-compile-file-form-defvar)
(byte-compile-file-form-defvar-function, byte-compile-lambda)
(byte-compile-defvar, byte-compile-file-form-defalias): Adjust callers.
This is for when lexical-binding is nil. The problem fixed was M-x
compile-defun leaving symbols with position in the compiled function's arglist
and interactive form. This fixes bug #55323. Also ensure the doc string is
correctly stripped when lexical-binding is t.
* lisp/emacs-lisp/bytecomp.el (byte-compile-lambda): For a (list ...)
interactive form when lexical-binding is nil, strip the positions from the
symbols in the form. Also strip the position from the symbols in the arglist.
(byte-compile-make-closure): (Twice) strip symbols from positions in the doc
string expression. Add comments.
* lisp/emacs-lisp/bytecomp.el
(byte-compile--suspicious-defcustom-choice): New function (bug#16271).
(byte-compile-nogroup-warn): Use it to warn about forms like
(choice (const :tag "foo" 'bar)).
These are symbols with position from source code, which should not be replaced
by bare symbols in, e.g., optimization functions.
* lisp/Makefile.in: (BYTE_COMPILE_FLAGS, compile-first case): Set
max-specpdl-size to 5000 for the benefit of lisp/emacs-lisp/comp.el.
* lisp/emacs-lisp/byte-opt.el (byte-optimize-form-code-walker)
(byte-optimize--rename-var, byte-optimize-if, byte-optimize-letX)
* lisp/emacs-lisp/bytecomp.el (byte-compile-recurse-toplevel)
(byte-compile-lambda)
* lisp/emacs-lisp/cconv.el (cconv-convert)
* lisp/emacs-lisp/macroexp.el (macroexp--expand-all): Preserve, e.g., (car
form) in the byte compiler, when this form's car is a symbol with position of
a special form, rather than replacing the symbol with a bare symbol, e.g.
'cond.
A supplementary commit to that on 2022-04-18:
* lisp/emacs-lisp/bytecomp.el (byte-compile--first-symbol-with-pos): Handle
vectors and records correctly.
* lisp/emacs-lisp/byte-run.el (byte-run--ssp-seen): Correct the doc string.
Correct the algorithm for determining the warning position to get the first
symbol-with-position in byte-compile--form-stack.
* lisp/emacs-lisp/bytecomp.el (byte-compile--first-symbol-with-pos): Function
renamed and amended from byte-compile--first-symbol.
(byte-compile--warning-source-offset): Call the new function above rather than
the old one.
This fixes bug #54990.
* lisp/emacs-lisp/bytecomp.el (byte-compile-lapcode): Remove positions from
symbols with positions in byte-switch tables, by temporarily removing the
entries from the table, and reinserting them amended.
This fixes bug #54079.
* lisp/emacs-lisp/bytecomp.el (byte-compile-initial-macro-environment): Change
the position of 'byte-run-strip-symbol-positions' in the eval-when-compile
entry. Add a call to `byte-run-strip-symbol-positions' in the
eval-and-compile entry.
Also increment a loop counter. This should fix bug #54248.
* lisp/emacs-lisp/byte-run.el (byte-run--strip-vector/record): increment the
loop counter 'i' in the main loop.
* lisp/emacs-lisp/bytecomp.el (byte-compile-out): call
byte-run-strip-symbol-positions on operands which are one-element lists.
This fixes bug #54098.
* lisp/emacs-lisp/byte-run.el (byte-run--strip-list)
(byte-run--strip-vector/record): New functions. These alter a list or
vector/record structure only where a symbol with position gets replaced by a
bare symbol.
(byte-run-strip-symbol-positions): Reformulate to use the two new functions.
(function-put): No longer strip positions from the second and third arguments.
* lisp/emacs-lisp/bytecomp.el (byte-compile-out): Remove the senseless
"stripping" of putative symbol positions from OPERAND, which is nil or a
number.
This happened with the tags of a condition-case. Also fix the detection of
circular lists while stripping the positions from symbols with position.
* lisp/emacs-lisp/byte-run.el (byte-run--circular-list-p): Remove.
(byte-run--strip-s-p-1): Write a value of t into a hash table for each cons or
vector/record encountered. (This is to prevent loops with circular
structures.) This is now done for all arguments, not just those detected as
circular lists.
* lisp/emacs-lisp/bytecomp.el (byte-compile-file-form-defvar)
(byte-compile-form, byte-compile-dynamic-variable-op)
(byte-compile-constant, byte-compile-push-constant): Remove redundant calls to
`bare-symbol'.
(byte-compile-lambda): call `byte-run-strip-symbol-positions' on the arglist.
(byte-compile-out): call `byte-run-strip-symbol-positions' on the operand.
This is the main call to this function in bytecomp.el.
* src/fns.c (hashfn_eq): Strip the position from an argument which is a symbol
with position.
(hash_lookup): No longer strip a position from a symbol with position.
(sxhash_obj): Add handling for symbols with position, substituting their bare
symbols when symbols with position are enabled.